Principles in Economics and Mathematics: the mathematical part

Principles in Economics and Mathematics: the mathematical part Bram De Rock Bram De Rock Mathematical principles 1/65

Practicalities about me Bram De Rock Office: R.42.6.218 E-mail: bderock@ulb.ac.be Phone: 02 650 4214 Mathematician who is doing research in Economics Homepage: http://www.revealedpreferences.org/bram.php Bram De Rock Mathematical principles 2/65

Practicalities about the course 12 hours on the mathematical part Micael Castanheira: 12 hours on the economics part Slides are available at MySBS and on http://mathecosolvay.com/spma/ Schedule Tuesday 17/9 and 24/9, 18.00-21.00, R42.2.107 Wednesday 18/9, 18.00-21.00, R42.2.103 Thursday 26/9, 18.00-21.00, R42.2.107 Course evaluation Written exam in the beginning of November to verify if you can apply the concepts discussed in class Compulsory for students in Financial Markets On a voluntary basis for students in Quantitative Finance Bram De Rock Mathematical principles 3/65

Course objectives and content Refresh some useful concepts needed in your other coursework No thorough or coherent study Interested student: see references for relevant material Content: 1 (derivatives, optimization, concavity) 2 (solving system of linear equations, matrices, linear (in)dependence) 3 Fundamentals on probability (probability and cumulative distributions, expectations of a random variable, correlation) Bram De Rock Mathematical principles 4/65

References Chiang, A.C. and K. Wainwright, Fundamental Methods of Mathematical Economics, Economic series, McGraw-Hill. Green, W.H., Econometric Analysis, Seventh Edition, Pearson Education limited. Luderer, B., V. Nollau and K. Vetters, Mathematical Formulas for Economists, Springer, New York. ULB-link Simon, C.P. and L. Blume Mathematics for Economists, Norton & Company, New York. Sydsaeter, K., A. Strom and P. Berck, Economists Mathematical Manual, Springer, New York. ULB-link Bram De Rock Mathematical principles 5/65

Outline Introduction 1 Introduction 2 3 4 Bram De Rock Mathematical principles 6/65

Role of functions Introduction = the study of functions Functions allow to exploit mathematical tools in Economics E.g. make consumption decisions max U(x 1, x 2 ) s.t. p 1 x 1 + p 2 x 2 = Y Characterization: x 1 = f (p 1, p 1, Y ) Econometrics: estimate f Allows to model/predict consumption behavior Warning about identification Causality: what is driving what? Functional structure: what is driving the result? Does the model allow to identify Bram De Rock Mathematical principles 7/65

Derivatives Introduction Marginal changes are important in Economics The impact of a infinitesimally small change of one of the variables Comparative statistics: what is the impact of a price change? : what is the optimal consumption bundle? Marginal changes are mostly studied by taking derivatives Characterizing the impact depends on the function f : D R n R k : (x 1,..., x n ) (y 1,... y k ) = f (x 1,..., x n ) We will always take k = 1 First look at n = 1 and then generalize Note: N Z Q R C Bram De Rock Mathematical principles 8/65

Outline Introduction 1 Introduction 2 3 4 Bram De Rock Mathematical principles 9/65

: f : D R R df dx = f f (x + x) f (x) = lim x 0 x Limit of quotient of differences If it exists, then it is called the derivative f is again a function E.g. f (x) = 3x 2 4 E.g. discontinuous functions, border of domain, f (x) = x Bram De Rock Mathematical principles 10/65

Some important derivatives and rules Let us abstract from specifying the domain D and assume that c, n R 0 If f (x) = c, then f (x) = 0 If f (x) = cx n, then f (x) = ncx n 1 If f (x) = ce x, then f (x) = ce x If f (x) = c ln(x), then f (x) = c 1 x (f (x) ± g(x)) = f (x) ± g (x) (f (x)g(x)) = f (x)g(x) + f (x)g (x) f (x)g (x) ( f (x) g(x) ) = f (x)g(x) f (x)g (x) g(x) 2 Bram De Rock Mathematical principles 11/65

Application: the link with marginal changes By definition it is the limit of changes Slope of the tangent line Increasing or decreasing function (and thus impact) Does the inverse function exist? First order approximation in some point c Based on expression for the tangent line in c f (c + x) f (c) + f (c)( x) More general approximation: Taylor expansion Bram De Rock Mathematical principles 12/65

Application: elasticities The elasticity of f in x: f (x)x f (x) The limit of the quotient of changes in terms of percentage Percentage change of the function: Percentage change of the variable: x Quotient: f (x+ x) f (x) x x f (x) Is a unit independent informative number E.g. the (price) elasticity of demand f (x+ x) f (x) f (x) x Bram De Rock Mathematical principles 13/65

Application: comparative statics for a simple market model Demand: Q = 10 4P Supply: Q = 2 + αp P = 8 4+α and Q = 8+10α 4+α dp dα = 8 (4+α) 2 and dq dα = 32 (4+α) 2 The elasticity of demand is 4P 10 4P Bram De Rock Mathematical principles 14/65

Some exercises Introduction Compute the derivative of the following functions (defined on R + ) f (x) = 17x 2 + 5x + 7 f (x) = x + 3 f (x) = 1 x 2 f (x) = 17x 2 e x f (x) = xln(x) x 2 4 Let f (x) : R R : x x 2 + 5x. Determine on which region f is increasing Is f invertible? Approximate f in 1 and derive an expression for the approximation error Compute the elasticity in 3 and 5 Bram De Rock Mathematical principles 15/65

The chain rule Introduction Often we have to combine functions If z = f (y) and y = g(x), then z = h(x) = f (g(x)) We have to be careful with the derivative A small change in x causes a chain reaction It changes y and this in turn changes z That is why dz dx = dz dy dy dx = f (y)g (x) Can easily be generalized to compositions of more than two functions dz dx = dz dy dy du dv dx E.g. if h(x) = e x 2, then h (x) = e x 2 2x Bram De Rock Mathematical principles 16/65

Higher order derivatives The derivative is again a function of which we can take derivatives Higher order derivatives describe the changes of the changes Notation f (x) or more generally f (n) (x) d dx ( df n d dx ) or more generally dx f (x) n E.g. if f (x) = 5x 3 + 2x, then f (x) = f (3) (x) = 30 Bram De Rock Mathematical principles 17/65

Application: concave and convex functions f : D R n R f is concave x, y D, λ [0, 1] : f (λx +(1 λ)y) λf (x)+(1 λ)f (y) If n = 1, x D : f (x) 0 f is convex x, y D, λ [0, 1] : f (λx +(1 λ)y) λf (x)+(1 λ)f (y) If n = 1, x D : f (x) 0 Bram De Rock Mathematical principles 18/65

Application: concave and convex functions Very popular and convenient assumptions in Economics E.g. optimization Sometimes intuitive interpretation E.g. risk-neutral, -loving, -averse Don t be confused with a convex set S is a set x, y S, λ [0, 1] : λx + (1 λ)y S Bram De Rock Mathematical principles 19/65

Some exercises Introduction Compute the first and second order derivative of the following functions (defined on R + ) f (x) = π f (x) = 5x + 3 f (x) = e 3x f (x) = ln(5x) f (x) = x 3 6x 2 + 17 Determine which of these functions are concave or convex Bram De Rock Mathematical principles 20/65

Outline Introduction 1 Introduction 2 3 4 Bram De Rock Mathematical principles 21/65

: f : D R n R Same applications in mind but now several variables E.g. what is the marginal impact of changing x 1, while controlling for other variables? Look at the partial impact: partial derivatives x i f (x 1,..., x n ) = f xi = lim xi 0 f (x 1,...,x i + x i,...,x n) f (x 1,...,x i,...,x n) x i Same interpretation as before, but now fixing remaining variables E.g. f (x 1, x 2, x 3 ) = 2x1 2x 2 5x 3 x 1 f (x 1, x 2, x 3 ) = 4x 1 x 2 x 2 f (x 1, x 2, x 3 ) = 2x1 2 x 3 f (x 1, x 2, x 3 ) = 5 Bram De Rock Mathematical principles 22/65

Partial derivative Introduction Geometric interpretation: slope of tangent line in the x i direction Same rules hold Higher order derivatives 2 xi 2 f (x 1,..., x n ) 2 x i x j f (x 1,..., x n ) = 2 x j x i f (x 1,..., x n ) E.g. 2 f (x x1 2 1, x 2, x 3 ) = 4x 2 and 2 x 1 x 3 f (x 1, x 2, x 3 ) = 0 Bram De Rock Mathematical principles 23/65

Some remarks Introduction Gradient: f (x 1,..., x n ) = ( f x 1,..., f x n ) Chain rule: special case x 1 = g 1 (t),..., x n = g n (t) and f (x 1,..., x n ) h(t) = f (x 1,..., x n ) = f (g 1 (t),..., g n (t)) dh(t) dt = h (t) = f (x 1,...,x n) dx 1 x 1 dt + + f (x 1,...,x n) x n E.g. f (x 1, x 2 ) = x 1 x 2, g 1 (t) = e t and g 2 (t) = t 2 h (t) = e t t 2 + e t 2t dx n dt Bram De Rock Mathematical principles 24/65

Some remarks Introduction Slope of indifference curve of f (x 1, x 2 ) Indifference curve: all (x 1, x 2 ) for which f (x 1, x 2 ) = C (with C some give number) Implicit function theorem: f (x 1, g(x 1 )) = C x 1 f (x 1, x 2 ) + x 2 f (x 1, x 2 ) dg dx 1 = 0 Slope = x f (x 1,x 2 ) 1 x f (x 1,x 2 ) 2 Bram De Rock Mathematical principles 25/65

Some exercises Introduction Compute the gradient and all second order partial derivatives for the following functions (defined on R + ) f (x 1, x 2 ) = x 2 1 2x 1x 2 + 3x 2 2 f (x 1, x 2 ) = ln(x 1 x 2 ) f (x 1, x 2, x 3 ) = e x 1+2x 2 3x1 x 3 Compute the marginal rate of substitution for the utility function U(x 1, x 2 ) = x α 1 x β 2 Bram De Rock Mathematical principles 26/65

Outline Introduction 1 Introduction 2 3 4 Bram De Rock Mathematical principles 27/65

: important use of derivatives Many models in economics entail optimizing behavior Maximize/Minimize objective subject to constraints Characterize the points that solve these models Note on Mathematics vs Economics Profit = Revenue - Cost Marginal revenue = marginal cost Marginal profit = zero Bram De Rock Mathematical principles 28/65

: formal problem max / min f (x 1,..., x n ) s.t. g 1 (x 1,..., x n ) = c 1 g m (x 1,..., x n ) = c m x 1,... x n 0 Inequality constraints are also possible Kuhn-Tucker conditions Bram De Rock Mathematical principles 29/65

Necessity and sufficiency Necessary conditions based on first order derivatives Local candidate for an optimum Sufficient conditions based on second order derivatives Necessary condition is sufficient if The constraints are convex functions E.g. no constraints, linear constraints,... The objective function is concave: global maximum is obtained The objective function is convex: global minimum is obtained Often the real motivation in Economics Bram De Rock Mathematical principles 30/65

Necessary conditions 1 Free optimization No constraints f (x ) = 0 if n = 1 x i f (x 1,..., x n ) = 0 for i = 1,..., n Intuitive given our geometric interpretation Bram De Rock Mathematical principles 31/65

Necessary conditions 2 with positivity constraints No g i constraints On the boundary extra optima are possible Often ignored: interior solutions xi 0 for i = 1,..., n xi x i f (x1,..., x n ) = 0 for i = 1,..., n x i f (x1,..., x n ) 0 for all i = 1,..., n simultaneously OR x i f (x1,..., x n ) 0 for all i = 1,..., n simultaneously Bram De Rock Mathematical principles 32/65

Necessary conditions 3 Constrained optimization without positivity constraints Define Lagrangian: L(x 1,..., x n, λ 1,..., λ m ) = f (x 1,..., x n ) λ 1 (g 1 (x 1,..., x n )) λ m (g m (x 1,..., x n )) x i L(x 1,..., x n, λ 1,..., λ m) = 0 for all i = 1,..., n λ j L(x 1,..., x n, λ 1,..., λ m) = 0 for all j = 1,..., m Alternatively: f (x 1,..., x n ) = λ 1 g 1(x 1,..., x n ) + + λ m g m (x 1,..., x n ) Some intuition: geometric interpretation Lagrange multiplier = shadow price Bram De Rock Mathematical principles 33/65

Application: utility maximization max U(x 1, x 2 ) = x α 1 x 1 α 2 s.t. p 1 x 1 + p 2 x 2 = Y L(x 1, x 2, λ 1 ) = x1 αx 1 α 2 λ 1 (p 1 x 1 + p 2 x 2 Y ) = αx α 1 1 x 1 α 2 λ 1 p 1 = 0 L x 1 L x 2 = (1 α)x α 1 x α 2 λ 1 p 2 = 0 L λ 1 = p 1 x 1 + p 2 x 2 Y = 0 x 1 = αy p 1, x 2 = (1 α)y p 2 and λ 1 = ( α p 1 ) α ( (1 α) p 2 ) (1 α) Bram De Rock Mathematical principles 34/65

Some exercises Introduction Find the optima for the following problems max / min x 3 12x 2 + 36x + 8 max / min x 3 1 x 3 2 + 9x 1x 2 min 2x 2 1 + x 1x 2 + 4x 2 2 + x 1x 3 + x 2 3 15x 1 max x 1 x 2 s.t. x 1 + 4x 2 = 16 max yz + xz s.t. y 2 + z 2 = 1 and xz = 3 Add positivity constraints to the above unconstrained problems and do the same Bram De Rock Mathematical principles 35/65

Outline Introduction Matrix algebra The link with vector spaces Application: solving a system of linear equations 1 Introduction 2 3 Matrix algebra The link with vector spaces Application: solving a system of linear equations 4 Bram De Rock Mathematical principles 36/65

Introduction Matrix algebra The link with vector spaces Application: solving a system of linear equations Matrices allow to formalize notation Useful in solving system of linear equations Useful in deriving estimators in econometrics Allows us to make the link with vector spaces Bram De Rock Mathematical principles 37/65

Outline Introduction Matrix algebra The link with vector spaces Application: solving a system of linear equations 1 Introduction 2 3 Matrix algebra The link with vector spaces Application: solving a system of linear equations 4 Bram De Rock Mathematical principles 38/65

Matrices Introduction Matrix algebra The link with vector spaces Application: solving a system of linear equations a 11 a 12 a 1m a 21 a 22 a 2m A = (a ij ) i=1,...,n;j=1,...,m =... a n1 a n2 a nm a ij R and A R n m n rows and m columns Square matrix if n = m Notable square matrices Symmetric matrix: a ij = a ji for all i, j = 1,..., n Diagonal matrix: a ij = 0 for all i, j = 1,..., n and i j Triangular matrix: only non-zero elements above (or below) the diagonal Bram De Rock Mathematical principles 39/65

Matrix manipulations Matrix algebra The link with vector spaces Application: solving a system of linear equations Let A, B R n m and k R Equality: A = B a ij = b ij for all i, j = 1,..., n Scalar multiplication: ka = (ka ij ) i=1,...,n;j=1,...,m Addition: A ± B = (a ij ± b ij ) i=1,...,n;j=1,...,m Dimensions must be equal Transposition: A = A t = (a ji ) j=1,...,m;i=1,...,n A R n m and A t R m n (A ± B) t = A t ± B t (ka) t = ka t (A t ) t = A Bram De Rock Mathematical principles 40/65

Matrix multiplication Matrix algebra The link with vector spaces Application: solving a system of linear equations Let A R n m and B R m k AB = ( m h=1 a ihb hj ) i=1,...,n;j=1,...,k Multiply the row vector of A with the column vector of B Aside: scalar/inner product and norm of vectors Orthogonal vectors Number of columns of A must be equal to number of rows of B AB BA, even if both are square matrices (AB) t = B t A t Bram De Rock Mathematical principles 41/65

Example Introduction Matrix algebra The link with vector spaces Application: solving a system of linear equations Let A = A t = ( ) 2 1, B = 1 1 ( ) 1 1 and C = 0 2 ( ) ( ) 2 1 1 0, B 1 1 t and C = 1 2 ( ) 2 0 AB = and BA = 1 1 ( ) 5 6 7 AC = 4 4 4 ( ) 2 1 (AB) t = B t A t = 0 1 ( ) 1 0 2 2 ( 1 2 ) 3 3 2 1 1 3 2 2 3 1 Bram De Rock Mathematical principles 42/65

Exercises Introduction Matrix algebra The link with vector spaces Application: solving a system of linear equations ( ) 3 6 Let A =, B = 1 2 ( ) 1 2 0 1 D = 4 0 3 1 Compute 3C, A + B, A D and D t Compute AB, BA, AC, CA, AD and DA ( ) 1 2, C = ( 1 1 ) and 3 1 Let A be a symmetric matrix, show then that A t = A A square matrix A is called idempotent if A 2 = A Verify which of the above matrices are idempotent Find the value ( of α that ) makes the following matrix 1 2 idempotent: α 2 Bram De Rock Mathematical principles 43/65

Matrix algebra The link with vector spaces Application: solving a system of linear equations Two numbers associated to square matrices: trace Let A, B, C R n n Trace(A) = tr(a) = n i=1 a ii Used in econometrics Properties tr(a t ) = tr(a) tr(a + B) = tr(a) + tr(b) tr(ca) = ctr(a) for any c R tr(ab) = tr(ba) tr(abc) = tr(bca) = tr(cab) tr(acb)(= tr(bac) = tr(cba)) Bram De Rock Mathematical principles 44/65

Matrix algebra The link with vector spaces Application: solving a system of linear equations Two numbers associated to square matrices: trace Example ( ) ( ) 3 6 1 2 Let A = and B = 1 2 3 1 ( ) ( ) 21 0 1 2 Then AB =, BA = and 7 0 10 20 ( ) 4 8 A + B = 2 3 tr(a) = 1, tr(b) = 0 and tr(a + B) = 1 tr(ab) = 21 = tr(ba) Bram De Rock Mathematical principles 45/65

Matrix algebra The link with vector spaces Application: solving a system of linear equations Two numbers associated to square matrices: determinant Let A R n n If n = 1, then det(a) = a 11 If n = 2, then det(a) = a 11 a 22 a 12 a 21 = a 11 det(a 22 ) a 12 det(a 21 ) ( ) a22 a If n = 3, then det(a) = a 11 det 23 a 32 a ( ) ( 33 ) a21 a a 12 det 23 a21 a + a a 31 a 13 det 22 33 a 31 a 32 Can be generalized to any n Works with columns too Bram De Rock Mathematical principles 46/65

Matrix algebra The link with vector spaces Application: solving a system of linear equations Two numbers associated to square matrices: determinant Let A, B R n n det(a t ) = det(a) det(a + B) det(a) + det(b) det(ca) = c det(a) for any c R det(ab) = det(ba) A is non-singular (or regular) if A 1 exists I.e. AA 1 = A 1 A = I n I n is a diagonal matrix with 1 on the diagonal Does not always exist det(a) 0 Bram De Rock Mathematical principles 47/65

Matrix algebra The link with vector spaces Application: solving a system of linear equations Two numbers associated to square matrices: determinant Example ( ) ( ) 3 6 1 2 Let A = and B = 1 2 3 1 ( ) ( ) 21 0 1 2 Then AB =, BA = and 7 0 10 20 ( ) 4 8 A + B = 2 3 det(a) = 0, det(b) = 7 and det(a + B) = 28 det(ab) = 0 = det(ba) Bram De Rock Mathematical principles 48/65

Exercises Introduction Matrix algebra The link with vector spaces Application: solving a system of linear equations 2 4 5 Let A = 0 3 0 1 0 1 Compute tr(a) and det( 2A) Show that for any triangular matrix A, we have that det(a) is equal to the product of the elements on the diagonal Let A, B R n n and assume that B is non-singular Show that tr(b 1 AB) = tr(a) Show that tr(b(b t B) 1 B t ) = n Let A, B R n n be two non-singular matrices Show that AB is then also invertible Give an expression for (AB) 1 Bram De Rock Mathematical principles 49/65

Outline Introduction Matrix algebra The link with vector spaces Application: solving a system of linear equations 1 Introduction 2 3 Matrix algebra The link with vector spaces Application: solving a system of linear equations 4 Bram De Rock Mathematical principles 50/65

A notion of vector spaces Matrix algebra The link with vector spaces Application: solving a system of linear equations A set of vectors V is a vector space if Addition of vectors is well-defined a, b V : a + b V Scalar multiplication is well-defined k R, a V : ka V We can take linear combinations k 1, k 2 R, a, b V : k 1 a + k 2 b V E.g. R 2 or more generally R n Counterexample R 2 + Bram De Rock Mathematical principles 51/65

Linear (in)dependence Matrix algebra The link with vector spaces Application: solving a system of linear equations Let V be a vector space A set of vectors v 1,..., v n V is linear dependent if one of the vectors can be written as a linear combination of the others k 1,... k n 1 R : v n = k 1 v 1 + + k n 1 v n 1 A set of vectors are linear independent if they are not linear dependent k 1,... k n R : k 1 v 1 + + k n v n = 0 k 1 = = k n = 0 In a vector space of dimension n, the number of linear independent vectors cannot be higher than n Bram De Rock Mathematical principles 52/65

Linear (in)dependence Matrix algebra The link with vector spaces Application: solving a system of linear equations Example R 2 is a vector space of dimension 2 v 1 = (1, 0), v 2 = (1, 2), v 3 = ( 1, 4) and v 4 = (2, 4) v 3 = 3v 1 + 2v 2, so v 1, v 2, v 3 are linear dependent v 4 = 2v 2, so v 2, v 4 are linear dependent v 1, v 2 are linear independent v 3 is linear independent Bram De Rock Mathematical principles 53/65

Link with matrices: rank Matrix algebra The link with vector spaces Application: solving a system of linear equations Let A R n m and B R m k The row rank of A is the maximal number of linear independent rows of A The column rank of A is the maximal number of linear independent columns of A Rank of A = column rank of A = row rank of A Properties rank(a) min(n, m) rank(ab) min(rank(a), rank(b)) rank(a) = rank(a t A) = rank(aa t ) If n = m, then A has maximal rank if and only if det(a) 0 Bram De Rock Mathematical principles 54/65

Link with matrices: rank Matrix algebra The link with vector spaces Application: solving a system of linear equations Example ( ) ( ) 3 6 1 2 Let A = and B = 1 2 3 1 ( ) ( ) 21 0 1 2 Then AB =, BA = 7 0 10 20 rank(a) = 1 and rank(b) = 2 rank(ab) = 1 Bram De Rock Mathematical principles 55/65

Exercises Introduction Matrix algebra The link with vector spaces Application: solving a system of linear equations 2 4 5 Let A = 0 3 0 1 0 1 Show in two ways that A has maximal rank Let A, B R n m Show that there need not be any relation between rank(a + B), rank(a) and rank(b) Show that if A is invertible, then it needs to have a maximal rank Bram De Rock Mathematical principles 56/65

Outline Introduction Matrix algebra The link with vector spaces Application: solving a system of linear equations 1 Introduction 2 3 Matrix algebra The link with vector spaces Application: solving a system of linear equations 4 Bram De Rock Mathematical principles 57/65

System of linear equations Matrix algebra The link with vector spaces Application: solving a system of linear equations Linear equations in the unknowns x 1,..., x m Not x 1 x 2, xm,... 2 Constraints hold with equality Not 2x 1 + 5x 2 3 { 2x1 + 3x E.g. 2 x 3 = 5 x 1 + 4x 2 + x 3 = 0 Bram De Rock Mathematical principles 58/65

Using matrix notation Matrix algebra The link with vector spaces Application: solving a system of linear equations m unknowns: x 1,..., x m n linear constraints: a i1 x 1 + + a im x m = b i with a i1,... a im, b i R and i = 1,..., n Ax = b A = (a ij ) i=1,...,n,j=1,...,m x = (x i ) i=1,...,n b = (b i ) i=1,...,n Homogeneous if b i = 0 for all i = 1,..., n Bram De Rock Mathematical principles 59/65

Solving these system Matrix algebra The link with vector spaces Application: solving a system of linear equations Solve this by logical reasoning Eliminate or substitute variables Can also be used for non-linear systems of equations Can be cumbersome for larger systems Use matrix notation Gaussian elimination of the augmented matrix (A b) Can be programmed Only for systems of linear equations Theoretical statements are possible Bram De Rock Mathematical principles 60/65

Example Introduction Matrix algebra The link with vector spaces Application: solving a system of linear equations { x1 + 4x 2 = 0 2x 1 + 3x 2 = 5 x 1 = 4x 2 11x 2 = 5 x 2 = 5 11 and x 1 = 20 11 ( ) ( ) ( ) 1 4 x1 0 = 2 3 5 x 2 Take linear combinations of rows of the augmented matrix Is ( the same as ) taking ( linear combinations ) ( of the equations ) 1 4 0 1 4 0 1 0 20 11 2 3 5 0 11 5 0 11 5 Bram De Rock Mathematical principles 61/65

Theoretical results Matrix algebra The link with vector spaces Application: solving a system of linear equations Consider the linear system Ax = b with A R n m This system has a solution if and only if rank(a) = rank(a b) rank(a) rank(a b) by definition Is the (column) vector b a linear combination of the column vectors of A? If rank(a) < rank(a B), the answer is no If rank(a) = rank(a B), the answer is yes The solution is unique if rank(a) = rank(a B) = m n m There are many solutions if rank(a) = rank(a B) < m n < m or too many constraints are redundant Bram De Rock Mathematical principles 62/65

Exercises Introduction Matrix algebra The link with vector spaces Application: solving a system of linear equations Solve the following systems of linear equations { x x1 + 2x 2 + 3x 3 = 1 1 x 2 + x 3 = 1 3x 1 + 2x 2 + x 3 = 1 and 3x 1 + x 2 + x 3 = 0 4x 1 + 2x 3 = 1 For which values of k does the following system of linear equations have a unique solution? { x1 + x 2 = 1 x 1 kx 2 = 1 Consider the linear system Ax = b with A R n n Show that this system has a unique solution if and only if A is invertible Give a formula for this unique solution Show that homogeneous systems of linear equations always have a (possibly non-unique) solution Bram De Rock Mathematical principles 63/65

Outline Introduction 1 Introduction 2 3 4 Bram De Rock Mathematical principles 64/65

To be continued Bram De Rock Mathematical principles 65/65